MaxView

XPK Start: Fri Apr 17 04:59:25 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-17 04:59:49.673754: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0417 04:59:49.850131 135544293873472 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 04:59:58.890663 135544293873472 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 04:59:58.893012 135544293873472 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-xt9l1-slice-job-0-0.mt-06-grad-accum-xt9l1:8482
I0417 05:00:00.052309 135544293873472 max_utils.py:284] Jax distributed system initialized!
I0417 05:00:06.163704 135544293873472 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 05:00:06.163805 135544293873472 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 05:00:06.163862 135544293873472 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0417 05:00:06.163902 135544293873472 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 05:00:06.836220 135544293873472 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 05:00:06.950053 135544293873472 checkpointing.py:688] Setting up checkpoint logger...
I0417 05:00:06.950165 135544293873472 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 05:00:06.950209 135544293873472 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 05:00:06.950408 135544293873472 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7b4625c50140>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 05:00:10.287077 135544293873472 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0417 05:00:10.287312 135544293873472 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b320c21bc20>}, handler_registry=None
I0417 05:00:10.287544 135544293873472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b320c21bc20>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0417 05:00:10.287593 135544293873472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b31f470f2f0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0417 05:00:10.287627 135544293873472 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b320c21bc20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b320c21bc20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b31f470f2f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b31f470f2f0>}).
I0417 05:00:10.287968 135544293873472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 05:00:10.288042 135544293873472 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b31f42b2020> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 05:00:11.757494 135544293873472 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints
I0417 05:00:12.612904 135544293873472 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b31f421be90>
I0417 05:00:12.613075 135544293873472 checkpointing.py:302] Checkpoint manager created!
I0417 05:00:14.337474 135544293873472 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 05:00:14.337591 135544293873472 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0417 05:00:15.762682 135544293873472 maxtext_utils.py:1805]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0417 05:00:15.762797 135544293873472 maxtext_utils.py:1805]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0417 05:00:15.762863 135544293873472 maxtext_utils.py:1805]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0417 05:00:15.762911 135544293873472 maxtext_utils.py:1805]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0417 05:00:15.762950 135544293873472 maxtext_utils.py:1805]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0417 05:00:15.762984 135544293873472 maxtext_utils.py:1805]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0417 05:00:15.763020 135544293873472 maxtext_utils.py:1805]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0417 05:00:15.763056 135544293873472 maxtext_utils.py:1805]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0417 05:00:15.763090 135544293873472 maxtext_utils.py:1805]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0417 05:00:15.763123 135544293873472 maxtext_utils.py:1805]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0417 05:00:15.763155 135544293873472 maxtext_utils.py:1805]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0417 05:00:15.763187 135544293873472 maxtext_utils.py:1805]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0417 05:00:16.150941 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0417 05:00:16.151028 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0417 05:00:16.165780 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0417 05:00:16.165850 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0417 05:00:16.194957 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0417 05:00:16.195030 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0417 05:00:16.209776 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0417 05:00:16.209828 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0417 05:00:16.239021 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0417 05:00:16.239085 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0417 05:00:16.253677 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0417 05:00:16.253739 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0417 05:00:16.297311 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0417 05:00:16.297377 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0417 05:00:16.311987 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0417 05:00:16.312043 135544293873472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0417 05:00:16.530168 135544293873472 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 05:00:16.530309 135544293873472 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 05:00:16.537005 135544293873472 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 05:00:16.537101 135544293873472 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 05:00:16.553574 135544293873472 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 05:00:16.553704 135544293873472 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 05:00:16.569483 135544293873472 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 05:00:16.569617 135544293873472 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 05:00:16.599429 135544293873472 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 05:00:16.599575 135544293873472 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 05:00:16.615547 135544293873472 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 05:00:16.615673 135544293873472 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 05:00:16.631444 135544293873472 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 05:00:16.631567 135544293873472 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 05:00:16.659716 135544293873472 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 05:00:16.659870 135544293873472 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 05:00:16.681063 135544293873472 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 05:00:16.681428 135544293873472 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 05:00:28.075878 135544293873472 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0417 05:00:28.084905 135544293873472 metric_logger.py:289] number parameters: 1.104 billion
I0417 05:00:30.001295 135544293873472 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0417 05:00:40.598572 135544293873472 checkpointing.py:798] Waited 10.597261905670166 seconds for step 0 to finish before starting checkpointing.
I0417 05:00:40.601144 135544293873472 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 05:00:40.602996 135544293873472 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0417 05:00:40.604511 135544293873472 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0.
I0417 05:00:41.327248 135544293873472 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 05:00:41.350225 135544293873472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0417 05:00:41.350300 135544293873472 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0417 05:00:41.617075 135416476075776 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0
I0417 05:00:41.620868 135544293873472 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.272168s
I0417 05:00:41.621640 135544293873472 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.384 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28651952743530273 s) (per-host)
I0417 05:00:41.621700 135544293873472 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286598s (batch_requests_ready=0.003484s, total_serialization_initiated=0.282438s, others=0.000676s)
I0417 05:00:41.621778 135544293873472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290545s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.290528)
I0417 05:00:41.622619 135544293873472 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0.
I0417 05:00:41.622961 135416484468480 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 05:20:41.622923
I0417 05:00:41.632645 135544293873472 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 05:00:41.632892 135414295459584 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 05:00:41.633009 135544293873472 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776402040.601126, 'wait_for_prev_duration_secs': 6.771087646484375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776402040.6030338, 'checkpointer_blocking_duration_secs': 1.0200979709625244, 'get_old_steps_start_time': 1776402041.6231587, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776402040.5989919, 'checkpoint_manager_blocking_duration_secs': 1.0339808464050293}
I0417 05:00:41.633113 135544293873472 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0417 05:00:41.633179 135544293873472 max_utils.py:750] 
Memstats: After params initialized:
I0417 05:00:41.633226 135544293873472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 05:00:41.633257 135544293873472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 05:00:41.633284 135544293873472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 05:00:41.633309 135544293873472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 05:00:42.056485 135544293873472 metric_logger.py:185] completed step: 0, seconds: 1.599, TFLOP/s/device: 33.981, Tokens/s/device: 5122.049, total_weights: 262144, loss: 10.876
I0417 05:00:42.059803 135544293873472 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/tensorboard/'
I0417 05:00:42.629267 135544293873472 metric_logger.py:185] completed step: 1, seconds: 12.052, TFLOP/s/device: 4.510, Tokens/s/device: 679.730, total_weights: 262144, loss: 10.876
I0417 05:00:43.207555 135544293873472 metric_logger.py:185] completed step: 2, seconds: 0.438, TFLOP/s/device: 124.175, Tokens/s/device: 18717.042, total_weights: 262144, loss: 10.561
I0417 05:00:43.786108 135544293873472 metric_logger.py:185] completed step: 3, seconds: 0.149, TFLOP/s/device: 364.501, Tokens/s/device: 54941.517, total_weights: 262144, loss: 10.270
I0417 05:00:44.142202 135410051184384 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776402043722719099, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776402043722719099, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 05:00:44.410273 135416459290368 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/items
I0417 05:00:45.969150    2586 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 05:00:48.616406 135414826157824 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/items/array_metadatas/process_0
I0417 05:01:23.867121 135414303852288 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.904812s (commit=36.273910s, array_metadata_write=1.630902s)
I0417 05:01:23.868293 135416484468480 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.138 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.533145904541016 s) (per-host)
I0417 05:01:23.868402 135416484468480 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.245329s.
I0417 05:01:24.960013 135416484468480 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776402043722719099, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 05:01:25.698194 135410051184384 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776402043722719099, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 05:01:26.078688 135416484468480 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000419s.
I0417 05:01:26.319772 135416484468480 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 05:01:27.249928 135416484468480 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.164612s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/items
I0417 05:01:27.251446 135416484468480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0/items
I0417 05:01:27.791819 135416484468480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0
I0417 05:01:29.198744 135416484468480 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0`.
I0417 05:01:29.199631 135416484468480 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.60 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/0
I0417 05:01:29.201164 135416484468480 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.578089s.
I0417 05:01:29.648378 135414295459584 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 05:01:29.648508 135414295459584 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 05:01:29.648575 135414295459584 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0417 05:01:29.650335 135414295459584 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0417 05:01:30.770418 135544293873472 metric_logger.py:185] completed step: 4, seconds: 0.579, TFLOP/s/device: 93.883, Tokens/s/device: 14151.074, total_weights: 262144, loss: 10.021
I0417 05:01:30.780649 135544293873472 metric_logger.py:185] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.017, Tokens/s/device: 14171.319, total_weights: 262144, loss: 9.819
I0417 05:01:31.355988 135544293873472 metric_logger.py:185] completed step: 6, seconds: 46.983, TFLOP/s/device: 1.157, Tokens/s/device: 174.362, total_weights: 262144, loss: 9.666
I0417 05:01:31.934286 135544293873472 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 6766.499, Tokens/s/device: 1019920.319, total_weights: 262144, loss: 9.560
I0417 05:01:32.512584 135544293873472 metric_logger.py:185] completed step: 8, seconds: 0.577, TFLOP/s/device: 94.242, Tokens/s/device: 14205.230, total_weights: 262144, loss: 9.495
I0417 05:01:33.091915 135544293873472 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0417 05:01:33.092582 135544293873472 checkpointing.py:798] Waited 0.0006825923919677734 seconds for step 9 to finish before starting checkpointing.
I0417 05:01:33.094566 135544293873472 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 05:01:33.096153 135544293873472 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0417 05:01:33.097701 135544293873472 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9.
I0417 05:01:33.810030 135544293873472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0417 05:01:33.810122 135544293873472 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0417 05:01:33.837027 135544293873472 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028361s
I0417 05:01:33.837771 135544293873472 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.632 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03705263137817383 s) (per-host)
I0417 05:01:33.837854 135544293873472 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037134s (batch_requests_ready=0.002850s, total_serialization_initiated=0.033592s, others=0.000692s)
I0417 05:01:33.837933 135544293873472 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041117s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.041099)
I0417 05:01:33.838578 135544293873472 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9.
I0417 05:01:33.838927 135415923193600 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 05:21:33.838880
I0417 05:01:33.849589 135544293873472 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 05:01:33.849889 135414295459584 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 05:01:33.850059 135544293873472 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776402093.094534, 'wait_for_prev_duration_secs': 7.224082946777344e-05, 'time_between_consecutive_saves_sec': 3.444092035293579, 'checkpointer_blocking_start_time': 1776402093.0961924, 'checkpointer_blocking_duration_secs': 0.7428927421569824, 'get_old_steps_start_time': 1776402093.8391094, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776402093.0928109, 'checkpoint_manager_blocking_duration_secs': 0.7572126388549805}
I0417 05:01:33.850176 135544293873472 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0417 05:01:33.850223 135544293873472 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 05:01:34.046760 135416484468480 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9
I0417 05:01:36.205068 135415956764416 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9/items
I0417 05:01:39.489052 135415939979008 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9/items/array_metadatas/process_0
I0417 05:02:19.277773 135415931586304 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.549968s (commit=40.323626s, array_metadata_write=1.226342s)
I0417 05:02:19.278975 135415923193600 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.733 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.478224992752075 s) (per-host)
I0417 05:02:19.279103 135415923193600 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.440055s.
I0417 05:02:21.517931 135415923193600 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 05:02:21.557898 135415923193600 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000384s.
I0417 05:02:22.578320 135415923193600 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.019768s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9/items
I0417 05:02:22.581236 135415923193600 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9/items
I0417 05:02:23.118834 135415923193600 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9
I0417 05:02:24.534052 135415923193600 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9`.
I0417 05:02:24.534942 135415923193600 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.44 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_06_grad_accum/checkpoints/9
I0417 05:02:24.536540 135415923193600 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.697490s.
I0417 05:02:24.536738 135414295459584 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 05:02:24.536871 135414295459584 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 05:02:24.536942 135414295459584 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0417 05:02:24.539212 135414295459584 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0417 05:02:24.539329 135544293873472 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0417 05:02:24.539443 135544293873472 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 05:02:24.540081 135544293873472 metric_logger.py:185] completed step: 9, seconds: 0.578, TFLOP/s/device: 94.017, Tokens/s/device: 14171.245, total_weights: 262144, loss: 9.457
Per train step:
 Total TFLOPs: 54.35 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 17 05:02:36 UTC 2026
EXIT_CODE=0